CN109816200A - Task method for pushing, device, computer equipment and storage medium - Google Patents
Task method for pushing, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN109816200A CN109816200A CN201811544266.7A CN201811544266A CN109816200A CN 109816200 A CN109816200 A CN 109816200A CN 201811544266 A CN201811544266 A CN 201811544266A CN 109816200 A CN109816200 A CN 109816200A
- Authority
- CN
- China
- Prior art keywords
- task
- user
- target user
- recognition
- face
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Collating Specific Patterns (AREA)
Abstract
This application involves field of artificial intelligence, a kind of task method for pushing, device, computer equipment and storage medium are provided.The described method includes: obtaining the recognition of face score and ID card information of the target user from the corresponding terminal of target user, the recognition of face score is that the terminal calls Ministry of Public Security's recognition of face interface to carry out the score that recognition of face obtains to target facial image；The corresponding similarity attenuation coefficient of the target user is calculated according to the ID card information；Based on the recognition of face score and the similarity attenuation coefficient using the user gradation assessment models trained, the current user's level of the target user is obtained；The current user's level screens the corresponding goal task of the target user from current task set, and the goal task is the task that task grade meets preset condition；The goal task is pushed into the terminal.Internet resources can be saved using the present processes.
Description
Technical field
This application involves field of artificial intelligence, more particularly to a kind of task method for pushing, device, computer equipment
And storage medium.
Background technique
With the rapid development of internet, there is crowdsourcing platform Internet-based, by crowdsourcing platform, lease is public
The enterprises such as department, finance company can be dispensed the tasks such as the work such as collection, investigation after leasing, providing a loan using internet.
Traditional crowdsourcing platform needs user to carry out recognition of face in registration and has come in order to ensure the authenticity of user
It can succeed in registration at authentication as long as recognition of face score is more than the preset threshold of unified setting, crowdsourcing platform is for note
Successfully user uniformly carries out task push to volume, however in order to guarantee the information security during task execution, partial task pair
The requirement of recognition of face score is higher than preset threshold when user's registration, and it is corresponding that the partial task is not achieved in recognition of face score
The user that recognition of face score requires can not apply for the partial task, therefore, carry out meeting when task push in the conventional mode
Lead to the waste of Internet resources.
Summary of the invention
Based on this, it is necessary in view of the above technical problems, provide a kind of task method for pushing that can save Internet resources,
Device, computer equipment and storage medium.
A kind of task method for pushing, which comprises
The recognition of face score and ID card information of the target user, the people are obtained from the corresponding terminal of target user
Face identification score is that the terminal calls Ministry of Public Security's recognition of face interface to carry out point that recognition of face obtains to target facial image
Number；
The corresponding similarity attenuation coefficient of the target user is calculated according to the ID card information；
The user gradation assessment models trained are used based on the recognition of face score and the similarity attenuation coefficient,
Obtain the current user's level of the target user；
The corresponding goal task of the target user, institute are screened from current task set according to the current user's level
Stating goal task is the task that task grade meets preset condition；
The goal task is pushed into the terminal.
Know in one of the embodiments, in the face for obtaining the target user from the corresponding terminal of target user
Before other score and ID card information, which comprises
Obtain the facial image to be detected of user to be detected；
Characteristics of image is extracted from the facial image to be detected；
Described image feature is inputted to the In vivo detection model trained, it is general to obtain the corresponding living body of the user to be detected
Rate predicted value；
When determining that the corresponding In vivo detection result of the user to be detected is first pre- according to the living body probabilistic forecasting value
If when result, Xiang Suoshu terminal sends confirmation instruction, the confirmation instruction is used to indicate the terminal for the face to be detected
Image is determined as target facial image；
When determining that the corresponding In vivo detection result of the user to be detected is second pre- according to the living body probabilistic forecasting value
If when result, Xiang Suoshu terminal sends warning message.
Described in being screened from current task set described according to the current user's level in one of the embodiments,
Before the corresponding goal task of target user, which comprises
Obtain the corresponding task identification of each task in the current task set；
The corresponding task type of each task and contract information are searched according to the task identification；
Each is obtained using the task grade evaluation model trained based on the task type and the contract information
It is engaged in corresponding task grade.
The generation step of the task grade evaluation model includes: in one of the embodiments,
The first training sample set is obtained, it includes historic task pair that first training sample, which concentrates each first training sample,
Task type, contract information and the first markup information answered；
Determine the model structure information of initiating task grade evaluation model, and the initialization initiating task grade assessment
The model parameter of model；
Based in first training sample task type and contract information using the initiating task grade assess mould
Type obtains the corresponding task grade of first training sample；
Based on the difference between obtained task grade and first markup information, the initiating task grade is adjusted
The model parameter of assessment models obtains goal task grade evaluation model；
The goal task grade evaluation model is determined as to the task grade evaluation model trained.
In one of the embodiments, the method also includes:
Task is completed in the history for obtaining the target user；
The corresponding task identification of task is completed according to the history and searches corresponding task scoring；
The current user's level of the target user is adjusted according to task scoring.
The generation step of the user gradation assessment models includes: in one of the embodiments,
The second training sample set is obtained, the second training sample of each of described second training sample concentration includes history target
The corresponding recognition of face score of user, ID card information and the second markup information；
Determine the model structure information of initial user grade evaluation model, and the initialization initial user grade assessment
The model parameter of model；
The corresponding similarity of the history target user is calculated according to the corresponding ID card information of the history target user
Attenuation coefficient；
The initial use is used based on the corresponding recognition of face score of the history target user and similarity attenuation coefficient
Family grade evaluation model obtains the corresponding user gradation of the history target user；
Based on the difference between the second markup information in obtained user gradation and the second training sample, described in adjustment
The model parameter of initial user grade evaluation model obtains target user's grade evaluation model；
Target user's grade evaluation model is determined as to the user gradation assessment models trained.
A kind of task driving means, described device include:
Data acquisition module, for obtained from the corresponding terminal of target user the target user recognition of face score and
ID card information, the recognition of face score are that the terminal calls Ministry of Public Security's recognition of face interface to carry out target facial image
The score that recognition of face obtains；
Similarity attenuation coefficient computing module, for calculating the corresponding phase of the target user according to the ID card information
Like degree attenuation coefficient；
Current user's level obtains module, for being used based on the recognition of face score and the similarity attenuation coefficient
The user gradation assessment models trained, obtain the current user's level of the target user；
Goal task screening module, for screening the target from current task set according to the current user's level
The corresponding goal task of user, the goal task are the task that task grade meets preset condition；
Goal task pushing module, for the goal task to be pushed to the terminal.
Described device in one of the embodiments, further include: In vivo detection module, for obtain user to be detected to
Detect facial image；Characteristics of image is extracted from the facial image to be detected；Described image feature is inputted to the work trained
Body detection model obtains the corresponding living body probabilistic forecasting value of the user to be detected；Sentence when according to the living body probabilistic forecasting value
When the corresponding In vivo detection result of the fixed user to be detected is the first default result, Xiang Suoshu terminal sends confirmation instruction, institute
It states and confirms that instruction is used to indicate the terminal and the facial image to be detected is determined as target facial image；When according to the work
When body probabilistic forecasting value determines that the corresponding In vivo detection result of the user to be detected is the second default result, Xiang Suoshu terminal hair
Send warning message.
A kind of computer equipment, including memory and processor, the memory are stored with computer program, the processing
Device realizes the step of task method for pushing described in above-mentioned any embodiment when executing the computer program.
A kind of computer readable storage medium, is stored thereon with computer program, and the computer program is held by processor
The step of task method for pushing described in above-mentioned any embodiment is realized when row.
Above-mentioned task method for pushing, device, computer equipment and storage medium, by being obtained from the corresponding terminal of target user
The recognition of face score and ID card information of the target user are taken, the target is then calculated according to the ID card information and is used
The corresponding similarity attenuation coefficient in family, and trained based on the recognition of face score and similarity attenuation coefficient use
User gradation assessment models obtain the current user's level of the target user, finally according to the current user's level from working as
The corresponding goal task of the target user is screened in preceding set of tasks, goal task is pushed into terminal, due to according to user
Grade screens task, reduces the data volume of server push, to save Internet resources.
Detailed description of the invention
Fig. 1 is the application scenario diagram of task method for pushing in one embodiment；
Fig. 2 is the flow diagram of task method for pushing in one embodiment；
Fig. 3 is the flow diagram of the generation step of task grade evaluation model in one embodiment；
Fig. 4 is the flow diagram of task method for pushing in another embodiment；
Fig. 5 is the structural block diagram of task driving means in one embodiment；
Fig. 6 is the internal structure chart of computer equipment in one embodiment.
Specific embodiment
It is with reference to the accompanying drawings and embodiments, right in order to which the objects, technical solutions and advantages of the application are more clearly understood
The application is further elaborated.It should be appreciated that specific embodiment described herein is only used to explain the application, not
For limiting the application.
Task method for pushing provided by the present application, can be applied in application environment as shown in Figure 1.Wherein, terminal 102
It is communicated by network with server 104.Call Ministry of Public Security's recognition of face interface to mesh in the corresponding terminal 102 of target user
After mark facial image progress recognition of face obtains recognition of face score, server 104 obtains recognition of face score and body from terminal
Part card information, the similarity attenuation coefficient of target user is then calculated according to ID card information, is decayed according to the similarity that must be beaten
Coefficient and recognition of face score obtain the current user's level of target user, so using the user gradation disaggregated model trained
Server obtains the corresponding task grade of each task in current task set afterwards, according to task grade from current task set
In filter out the meeting preset condition of the task, filtering out for task is determined as goal task, finally pushes to goal task
Terminal 102.
Wherein, terminal 102 can be, but not limited to be various personal computers, laptop, smart phone, tablet computer
With portable wearable device, server 104 can use the server set of the either multiple server compositions of independent server
Group realizes.
In one embodiment, as shown in Fig. 2, providing a kind of task method for pushing, it is applied in Fig. 1 in this way
It is illustrated for server, comprising the following steps:
Step S202 obtains the recognition of face score and ID card information of target user from the corresponding terminal of target user.
Wherein, recognition of face score is that terminal calls Ministry of Public Security's recognition of face interface to carry out face knowledge to target facial image
The score not obtained, target user refer to the user by recognition of face, i.e. recognition of face score is more than preset threshold
User；ID card information refers to the information obtained according to the identity card of target user, name, identity card including target user
Number, date of birth etc.；Target facial image refers to the image that can be used as recognition of face after In vivo detection passes through.
Specifically, ID card information can be the information that user inputs in terminal interface, is also possible to terminal and passes through acquisition
ID Card Image, and ID card information of the Text region to obtain is carried out to ID Card Image by OCR technique.Implement at one
In example, terminal carries out In vivo detection after getting ID card information, to user, and after In vivo detection passes through, terminal needs
(such as 3S) collects the face image of user in the preset period, using collected face image as target facial image,
It does not collect face image yet when the preset time is exceeded, then restarts to carry out In vivo detection, further, terminal calls the Ministry of Public Security
Recognition of face interface to target facial image carry out recognition of face, terminal when calling Ministry of Public Security's recognition of face interface, upload
Interface parameters includes ID card No. and target facial image, and Ministry of Public Security's recognition of face interface can be searched according to ID card No.
To corresponding identity card picture, the identity card picture found is then subjected to face characteristic with target facial image and is compared, is obtained
To recognition of face score, after terminal gets the recognition of face score of Ministry of Public Security's recognition of face interface return, judge that the face is known
Whether other score is greater than preset threshold, if being greater than, illustrates that the recognition of face of the user passes through, which is determined as target
User, and the recognition of face score and ID card information of the target user are sent to server.
Step S204 calculates the corresponding similarity attenuation coefficient of target user according to ID card information.
Wherein, current face's image that similarity attenuation coefficient is used to characterize user declines with identity card facial image similarity
A possibility that subtracting size, because over time, the appearance of people usually has certain variation, this variation is equally embodied in
In facial image similarity, therefore with the increase of identity card service life, current face's image and the identity card of user is handled
When the identity card facial image that acquires between similarity can decay therewith, and under normal conditions, in service life identical feelings
Under condition, a possibility that age smaller user, similarity decays, is bigger.
In one embodiment, process, definable similarity attenuation coefficient are the use duration of identity card to simplify the calculation
The ratio in (time) and age, if the service life of identity card is in November, 2008 28-2018 November 28, the date of birth is
On January 1st, 1998, then similarity attenuation coefficient is (2018-2008)/(2018-1998).It is appreciated that in other embodiments
In, similarity attenuation coefficient can also be calculated by other means, as long as the current face's image and body of user can be characterized
A possibility that part witness's face image similarity decays size.
Step S206 assesses mould using the user gradation trained based on recognition of face score and similarity attenuation coefficient
Type obtains the current user's level of target user.
Specifically, user gradation assessment models can pass through engineering for assessing the user gradation of target user
The mode training of habit obtains.User gradation is used to characterize the confidence level of user's recognition of face, and user gradation is higher, can indicate user
The confidence level of recognition of face is higher.The setting of specific user gradation can be set in advance by technical staff according to business demand
It is fixed, such as may be set to high, medium and low, it also can be set as level-one, second level, three-level ... ..., n grades.
In the present embodiment, it is contemplated that the influence that human face similarity degree is decayed to recognition of face score can decline according to similarity
Subtract coefficient and recognition of face score determines the user gradation of target user jointly.Server is in the face for getting target user
After identifying score and similarity attenuation coefficient, based on these data using the user gradation assessment models trained to target user
User gradation assessed, to obtain the current user's level of target user.
Step S208 screens the corresponding goal task of target user according to current user's level from current task set.
Wherein, current task set refers to the set of all tasks that can currently push composition, and goal task is task
Grade meets the task of preset condition, and preset condition includes that task grade is identical as the current user's level of target user and appoint
Grade of being engaged in is lower than at least one of the current task grade of target user.For example, as the delimitation of user gradation be it is high,
In, it is low, then when the user gradation of target user be it is rudimentary when, corresponding goal task be rudimentary task, work as user gradation
When being advanced, corresponding goal task can be the task of all grades.
In the present embodiment, server needs to determine the corresponding task grade of each task in current task set first, this
Sample is after having determined the current user's level of target user, so that it may according to the current user's level of target user from current task
The meeting preset condition of the task is filtered out in set, these tasks are determined as to the goal task of the target user.It is understood that
, in the present embodiment, the delimitation of task grade and the delimitation of user gradation are identical, for example, user gradation delimit for it is high, in,
Low, then task grade equally delimited to be high, medium and low, for another example, user gradation and task grade can also all delimit for level-one,
Second level, three-level, level Four etc..
Goal task is pushed to terminal by step S210.
Specifically, server, which can be, pushes to terminal for goal task after the task acquisition request for receiving terminal；
It is also possible to according to the preset period, such as every other day, pushes a goal task to terminal, specifically when push, it can root
According to being previously set, the application is it is not limited here.
In above-mentioned task method for pushing, server is known by the face for obtaining target user from the corresponding terminal of target user
Then other score and ID card information calculate the corresponding similarity attenuation coefficient of target user according to ID card information, and are based on
Recognition of face score and similarity attenuation coefficient obtain the current use of target user using the user gradation assessment models trained
Family grade finally screens the corresponding goal task of target user, by target according to current user's level from current task set
Task pushes to terminal, in the application, due to being screened according to user gradation to task, on the one hand reduces server push
The data volume sent saves Internet resources, on the other hand, since the task specific aim of push is higher, can save user's application
Selection time when task, to promote the efficiency of task application.
In one embodiment, in the recognition of face score and identity for obtaining target user from the corresponding terminal of target user
Before card information, comprising: obtain the facial image to be detected of user to be detected；It is special that image is extracted from facial image to be detected
Sign；Characteristics of image is inputted to the In vivo detection model trained, obtains the corresponding living body probabilistic forecasting value of user to be detected；Work as root
When determining that the corresponding In vivo detection result of user to be detected is the first default result according to living body probabilistic forecasting value, sent to terminal true
Recognize instruction, confirmation instruction is used to indicate terminal and facial image to be detected is determined as target facial image；When according to living body probability
When predicted value determines that the corresponding In vivo detection result of user to be detected is the second default result, warning message is sent to terminal.
Wherein, user to be detected refers to the user for needing to carry out In vivo detection；Facial image to be detected refers to be checked
Survey the facial image of the user to be detected of the terminal acquisition of user；In vivo detection model is to be detected for being obtained according to characteristics of image
The In vivo detection of user is as a result, In vivo detection result includes the first default result and the second default result, wherein the first default knot
It is living body that fruit, which characterizes user to be detected, and it is non-living body that the second default result, which characterizes user to be detected,.
In the present embodiment, server is after getting the facial image to be detected of user to be detected, from face figure to be detected
Characteristics of image is extracted as in, then characteristics of image is input in the In vivo detection model trained, obtains user's to be detected
Living body probabilistic forecasting value.When server determines user to be detected for living body according to living body probabilistic forecasting value, sent to terminal true
Recognize instruction, facial image to be detected is determined as target facial image, then calls public security by terminal after receiving confirmation instruction
Recognition of face interface in portion's carries out recognition of face to the target facial image and obtains the face of Ministry of Public Security's recognition of face interface return
It identifies score, when the recognition of face score of return is more than preset threshold, shows that user to be detected has passed through recognition of face, terminal
The user to be detected is determined as target user；When server determines that user to be detected is non-living body according to living body probabilistic forecasting value
When, warning message is sent to terminal, terminal can prompt user to re-start man face image acquiring after receiving warning message,
And the facial image of acquisition is again sent to server and carries out In vivo detection, when the number of In vivo detection is more than preset times
User setting to be detected can be black list user by (such as 3 times), and for black list user, server will be pushed away without task
It send.
In one embodiment, can training image Feature Selection Model in advance, when extracting characteristics of image, by people to be detected
Face image is input in image characteristics extraction model, to obtain characteristics of image.In another embodiment, DoG can be used
(difference of Gaussian) filter is filtered pretreatment to image to be detected and obtains intermediate frequency information therein,
Fourier transform feature is extracted from pretreated two dimensional image using Fourier transformer as characteristics of image.
In one embodiment, Feature Selection Model can be trained in the following manner and be obtained: obtain face image set；Really
Determine initial characteristics to extract the model structure information of model and be initially generated the network structure information of confrontation network, and initialization should
The model parameter that initial characteristics extract model is initially generated the network parameter of confrontation network with this；For in face image set
Facial image is executed to lower ginseng step: facial image input initial characteristics being extracted model, are obtained and the facial image pair
The characteristics of image answered；Obtained characteristics of image is inputted into initial generator, obtains generating facial image；Based on obtained life
At the similarity between facial image and the living body faces image, the ginseng that initial characteristics extract model and initial generator is adjusted
Number；Initial characteristics adjusted extraction model is determined as Feature Selection Model.Wherein, above-mentioned initial generator is to be initially generated
The generator in network is fought, being initially generated confrontation network can be in order to which training characteristics extract model and predetermined include
The generation of initial generator and initial arbiter fights network (GAN, Generative Adversarial Networks),
In, initial generator is for generating image, and initial arbiter is for determining that inputted image is to generate image or true figure
Picture.
In one embodiment, In vivo detection model can be trained in the following manner and be obtained；Obtain training sample set, instruction
Practicing each training sample in sample set includes sample facial image and corresponding living body probability value；Determine initial In vivo detection mould
The model structure information of type, and the model parameter of the initial In vivo detection model of initialization；From the sample face in training sample
Image zooming-out sample image feature；Sample image feature is inputted into initial In vivo detection model, it is corresponding to obtain sample facial image
Sample living body probabilistic forecasting value；Based on the difference between the living body probability value in sample living body probabilistic forecasting value and training sample
It is different, the model parameter of initial In vivo detection model is adjusted, target In vivo detection model is obtained；Target In vivo detection model is determined
For the In vivo detection model trained.
In one embodiment, the corresponding mesh of target user is being screened from current task set according to current user's level
Before mark task, the above method further include: obtain the corresponding task identification of each task in current task set；According to task mark
Know and searches the corresponding task type of each task and contract information；Task based access control type and contract information are using having trained for task
Grade evaluation model obtains the corresponding task grade of each task.
Specifically, task type refers to classification belonging to task, after including but not limited to logistics survey tasks, factoring are rented
Prospecting task, assets reconnoitre task, wherein logistics survey tasks refer to that the scene for business rental object arrival is veritified and appoint
Business, prospecting task refers to the acquisition prospecting task of live business circumstance after renting for factoring enterprise, assets prospecting after factoring is rented
Task refers to the regular on-the-spot make an inspection tour task of assets after renting for enterprise；Contract information refers to that business corresponding with task is closed
Same relevant information, business contract for example can be the contract of lease of property, and contract information includes subject matter of a contract assets, the case-involving amount of money of contract
Deng subject matter of a contract assets refer to object pointed by the rights or obligation in the corresponding contract documents of task, the case-involving gold of contract
Volume for example can be cash pledge, guarantee fund, rent etc..Task grade evaluation model is for assessing the grade of task.Task
Mark is used for some task of unique identification, can be by number, letter or combinations thereof, the corresponding task type of each task and conjunction
With information storage corresponding with the task identification of the task.
In the present embodiment, server obtains the task identification of each task in current task set first, and then basis is appointed
Business mark searches the corresponding task type of each task and contract information from database, and task type and contract information are inputted
The corresponding task grade of each task can be obtained into the task grade evaluation model trained.
In above-described embodiment, assessed by obtaining the corresponding task type of task and contract information, and using task grade
Model obtains the corresponding task grade of task, and efficiency and accuracy that task grade determines can be improved.
In one embodiment, as shown in figure 3, the generation step of task grade evaluation model includes:
Step S302 obtains the first training sample set, and it includes history that the first training sample, which concentrates each first training sample,
The corresponding task type of task, contract information and the first markup information.
Wherein, the first markup information is for characterizing the corresponding task grade of historic task.In one embodiment, the first mark
Note information can be the vector comprising task class letter, for example, using vector when task grade includes high, medium and low three-level
(1,0,0) advanced tasks are characterized, intermediate task is characterized with vector (0,1,0), characterizes low-level tasks with vector (0,0,1)；Another
In one embodiment, the first markup information can be the vector including the first probability, the second probability and third probability, wherein
First probability is for characterizing a possibility that historic task is advanced tasks, and the second probability is for characterizing historic task as intermediate task
A possibility that, third probability is for characterizing a possibility that historic task is low-level tasks.
Step S304, determines the model structure information of initiating task grade evaluation model, and initializes the initiating task
The model parameter of grade evaluation model.
Specifically, initiating task grade evaluation model can be the various machine learning models that classification feature may be implemented,
For different types of model, the model structure information of required determination is not also identical.For example, task grade evaluation model can be with
For decision tree, logistic regression, naive Bayesian, neural network etc..
Further, it is possible to which each model parameter of task grade evaluation model is carried out just with some different small random numbers
Beginningization." small random number " is used to guarantee that model will not enter saturation state because weight is excessive, so as to cause failure to train, " no
It is used to together " guarantee that model can normally learn.
Step S306, based in the first training sample task type and contract information using initiating task grade assess mould
Type obtains the corresponding task grade of the first training sample.
Specifically, the corresponding task type of historic task and contract information can be mapped as input vector, will input to
In amount input initiating task grade evaluation model, so as to obtain the task grade of historic task in the first training sample.
Step S308 adjusts initiating task etc. based on the difference between obtained task grade and the first markup information
The model parameter of grade assessment models, obtains goal task grade evaluation model.
Goal task grade evaluation model is determined as the task grade evaluation model trained by step S310.
Specifically, it can use preset loss function (for example, L1 norm or L2 norm etc.) and calculate obtained go through
The difference between the first markup information in the task grade and training sample of history task, and based on the resulting difference tune of calculating
The model parameter of whole above-mentioned initial In vivo detection model, and when meeting default training termination condition, obtain target user's grade
Assessment models, wherein default training termination condition includes but is not limited to: the training time is more than preset threshold；Frequency of training is more than pre-
If number；It calculates obtained difference and is less than default discrepancy threshold.It, can be using various implementations based in the present embodiment
Calculate the model parameter of the above-mentioned user gradation assessment models of resulting discrepancy adjustment.For example, BP (Back Propagation, reversely
Propagate) algorithm or SGD (Stochastic Gradient Descent, stochastic gradient descent) algorithm.Further, by target
User gradation assessment models are determined as the user gradation assessment models trained.
In one embodiment, the above method further include: task is completed in the history for obtaining target user；According to history
The corresponding task identification of completion task searches corresponding task scoring；It is scored according to task and adjusts the active user etc. of target user
Grade.
Specifically, task is completed in the history that server obtains target user, is completed task corresponding according to history
The corresponding task of identifier lookup of being engaged in scores, and is scored according to task and adjusts the current user's level of target user.Wherein, task scores
For characterizing task publisher to the satisfaction of task performance.
In one embodiment, when the task quantity that user completes is more than first threshold and the corresponding task scoring of task
When mean value is more than second threshold, the grade of active user can be promoted；When the task quantity that target user completes is more than
When the mean value that first threshold and the corresponding task of task score is less than third threshold value, the corresponding grade of user is reduced.
In above-described embodiment, the current user's level of target user is adjusted by task scoring, use can be excited
The enthusiasm of family completion task.
In one embodiment, the generation step of above-mentioned user gradation assessment models includes: to obtain the second training sample set,
The second training sample of each of second training sample concentration includes the corresponding recognition of face score of history target user, identity card letter
Breath and the second markup information；Determine model structure information of initial user grade evaluation model, and initialization initial user etc.
The model parameter of grade assessment models；The corresponding phase of history target user is calculated according to the corresponding ID card information of history target user
Like degree attenuation coefficient；Based on the corresponding recognition of face score of history target user and similarity attenuation coefficient using initial user etc.
Grade assessment models, obtain the corresponding user gradation of history target user；Based on obtained user gradation and the second training sample
In the second markup information between difference, adjust initial user grade evaluation model model parameter, obtain target user etc.
Grade assessment models；Target user's grade evaluation model is determined as to the user gradation assessment models trained.
Wherein, the second markup information is for characterizing the corresponding user gradation of history target user.In one embodiment,
Two markup informations can be the vector comprising user class identifier, for example, when user gradation includes high, medium and low three-level, with to
It measures (1,0,0) and characterizes advanced level user, characterize intermediate users with vector (0,1,0), characterize less advanced users with vector (0,0,1)；?
In another embodiment, the second markup information can be the vector including the first probability, the second probability and third probability,
In, the first probability is for characterizing a possibility that history target user is advanced level user, and the second probability is for characterizing history target use
A possibility that family is intermediate users, third probability is for characterizing a possibility that history target user is less advanced users.
It is referred in the application in other embodiments it is appreciated that closing other explanations in this present embodiment and limiting
Description, this will not be repeated here by the application.
In one embodiment, as shown in figure 4, providing a kind of task method for pushing, comprising the following steps:
Step S402, terminal obtains the facial image and ID card information to be detected of user to be detected, by face to be detected
Image is sent to server；
Step S404 after server receives facial image to be detected, extracts characteristics of image from facial image to be detected；
Characteristics of image is inputted the In vivo detection model trained by step S406, server, and it is corresponding to obtain user to be detected
Living body probabilistic forecasting value；
Step S408, when determining that the corresponding In vivo detection result of user to be detected is living body according to living body probabilistic forecasting value
When, server sends confirmation instruction to terminal；
Facial image to be detected is determined as target facial image after receiving confirmation instruction by step S410, terminal；
Step S412, terminal calls Ministry of Public Security's recognition of face interface to carry out recognition of face to target facial image, and obtains
The recognition of face score that Ministry of Public Security's recognition of face interface returns will be to when recognition of face score is more than corresponding preset threshold
Detection user is determined as target user；
Step S414, server obtain the recognition of face score and ID card information of target user from terminal；
Step S416, server calculate the corresponding similarity attenuation coefficient of target user according to ID card information；
Step S418, server are commented based on recognition of face score and similarity attenuation coefficient using the user gradation trained
Estimate model, obtains the current user's level of target user；
Step S420, server obtains the corresponding task identification of each task in current task set, according to task identification
The corresponding task type of each task and contract information are searched, task based access control type and contract information are using the task dispatching trained
Grade assessment models obtain the corresponding task grade of each task；
Step S422, it is default that server filters out task grade satisfaction according to current user's level from current task set
Goal task is pushed to terminal by the goal task of condition.
It should be understood that although each step in the flow chart of Fig. 2-4 is successively shown according to the instruction of arrow,
These steps are not that the inevitable sequence according to arrow instruction successively executes.Unless expressly stating otherwise herein, these steps
Execution there is no stringent sequences to limit, these steps can execute in other order.Moreover, at least one in Fig. 2-4
Part steps may include that perhaps these sub-steps of multiple stages or stage are not necessarily in synchronization to multiple sub-steps
Completion is executed, but can be executed at different times, the execution sequence in these sub-steps or stage is also not necessarily successively
It carries out, but can be at least part of the sub-step or stage of other steps or other steps in turn or alternately
It executes.
In one embodiment, as shown in figure 5, providing a kind of task driving means 500, comprising:
Data acquisition module 502, for obtained from the corresponding terminal of target user target user recognition of face score and
ID card information, recognition of face score are that terminal calls Ministry of Public Security's recognition of face interface to carry out recognition of face to target facial image
Obtained score；
Similarity attenuation coefficient computing module 504, for calculating the corresponding similarity of target user according to ID card information
Attenuation coefficient；
Current user's level obtains module 506, has instructed for being used based on recognition of face score and similarity attenuation coefficient
Experienced user gradation assessment models, obtain the current user's level of target user；
Goal task screening module 508, for screening target user from current task set according to current user's level
Corresponding goal task, goal task are the task that task grade meets preset condition；
Goal task pushing module 510, for goal task to be pushed to terminal.
In one embodiment, above-mentioned apparatus further include: In vivo detection module, for obtaining the to be detected of user to be detected
Facial image；Characteristics of image is extracted from facial image to be detected；Characteristics of image is inputted to the In vivo detection model trained, is obtained
To the corresponding living body probabilistic forecasting value of user to be detected；When according to the corresponding living body of living body probabilistic forecasting value judgement user to be detected
When testing result is the first default result, confirmation instruction is sent to terminal, confirmation instruction is used to indicate terminal for face to be detected
Image is determined as target facial image；When determining that the corresponding In vivo detection result of user to be detected is according to living body probabilistic forecasting value
When the second default result, warning message is sent to terminal.
In one embodiment, above-mentioned apparatus further include: task level determination module, for obtaining in current task set
The corresponding task identification of each task；The corresponding task type of each task and contract information are searched according to task identification；It is based on
Task type and contract information use the task grade evaluation model trained to obtain the corresponding task grade of each task.
In one embodiment, above-mentioned apparatus further include: task grade evaluation model generation module, for obtaining the first instruction
Practice sample set, it includes the corresponding task type of historic task, contract information that the first training sample, which concentrates each first training sample,
And first markup information；Determine the model structure information of initiating task grade evaluation model, and initialize the initiating task etc.
The model parameter of grade assessment models；Based in the first training sample task type and contract information commented using initiating task grade
Estimate model and obtains the corresponding task grade of the first training sample；Based between obtained task grade and the first markup information
Difference adjusts the model parameter of initiating task grade evaluation model, obtains goal task grade evaluation model；By goal task etc.
Grade assessment models are determined as the task grade evaluation model trained.
In one embodiment, above-mentioned apparatus further include: user gradation adjusts module, for obtaining the history of target user
Task is completed；The corresponding task identification of task is completed according to history and searches corresponding task scoring；It is scored and is adjusted according to task
The current user's level of whole target user.
In one embodiment, above-mentioned apparatus further include: user gradation assessment models generation module, for obtaining the second instruction
Practice sample set, the second training sample of each of the second training sample concentration includes the corresponding recognition of face point of history target user
Number, ID card information and the second markup information；Determine the model structure information of initial user grade evaluation model, and initialization
The model parameter of initial user grade evaluation model；History target is calculated according to the corresponding ID card information of history target user to use
The corresponding similarity attenuation coefficient in family；It is used based on the corresponding recognition of face score of history target user and similarity attenuation coefficient
Initial user grade evaluation model obtains the corresponding user gradation of history target user；Based on obtained user gradation and
The difference between the second markup information in two training samples adjusts the model parameter of initial user grade evaluation model, obtains
Target user's grade evaluation model；Target user's grade evaluation model is determined as to the user gradation assessment models trained.
Specific about task driving means limits the restriction that may refer to above for task method for pushing, herein not
It repeats again.Modules in above-mentioned task driving means can be realized fully or partially through software, hardware and combinations thereof.On
Stating each module can be embedded in the form of hardware or independently of in the processor in computer equipment, can also store in a software form
In memory in computer equipment, the corresponding operation of the above modules is executed in order to which processor calls.
In one embodiment, a kind of computer equipment is provided, which can be server, internal junction
Composition can be as shown in Figure 6.The computer equipment include by system bus connect processor, memory, network interface and
Database.Wherein, the processor of the computer equipment is for providing calculating and control ability.The memory packet of the computer equipment
Include non-volatile memory medium, built-in storage.The non-volatile memory medium is stored with operating system, computer program and data
Library.The built-in storage provides environment for the operation of operating system and computer program in non-volatile memory medium.The calculating
The database of machine equipment is used for the related data of store tasks method for pushing.The network interface of the computer equipment is used for and outside
Terminal by network connection communication.To realize a kind of task method for pushing when the computer program is executed by processor.
It will be understood by those skilled in the art that structure shown in Fig. 6, only part relevant to application scheme is tied
The block diagram of structure does not constitute the restriction for the computer equipment being applied thereon to application scheme, specific computer equipment
It may include perhaps combining certain components or with different component layouts than more or fewer components as shown in the figure.
In one embodiment, a kind of computer equipment, including memory and processor are provided, which is stored with
Computer program, the processor realize the task push side provided in any one embodiment of the application when executing computer program
The step of method.
In one embodiment, a kind of computer readable storage medium is provided, computer program is stored thereon with, is calculated
Machine program realizes the step of task method for pushing provided in any one embodiment of the application when being executed by processor.
Those of ordinary skill in the art will appreciate that realizing all or part of the process in above-described embodiment method, being can be with
Relevant hardware is instructed to complete by computer program, the computer program can be stored in a non-volatile computer
In read/write memory medium, the computer program is when being executed, it may include such as the process of the embodiment of above-mentioned each method.Wherein,
To any reference of memory, storage, database or other media used in each embodiment provided herein,
Including non-volatile and/or volatile memory.Nonvolatile memory may include read-only memory (ROM), programming ROM
(PROM), electrically programmable ROM (EPROM), electrically erasable ROM (EEPROM) or flash memory.Volatile memory may include
Random access memory (RAM) or external cache.By way of illustration and not limitation, RAM is available in many forms,
Such as static state RAM (SRAM), dynamic ram (DRAM), synchronous dram (SDRAM), double data rate sdram (DDRSDRAM), enhancing
Type SDRAM (ESDRAM), synchronization link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM
(RDRAM), direct memory bus dynamic ram (DRDRAM) and memory bus dynamic ram (RDRAM) etc..
Each technical characteristic of above embodiments can be combined arbitrarily, for simplicity of description, not to above-described embodiment
In each technical characteristic it is all possible combination be all described, as long as however, the combination of these technical characteristics be not present lance
Shield all should be considered as described in this specification.
The several embodiments of the application above described embodiment only expresses, the description thereof is more specific and detailed, but simultaneously
It cannot therefore be construed as limiting the scope of the patent.It should be pointed out that coming for those of ordinary skill in the art
It says, without departing from the concept of this application, various modifications and improvements can be made, these belong to the protection of the application
Range.Therefore, the scope of protection shall be subject to the appended claims for the application patent.
Claims (10)
1. a kind of task method for pushing, which comprises
The recognition of face score and ID card information of the target user are obtained from the corresponding terminal of target user, the face is known
Other score is that the terminal calls Ministry of Public Security's recognition of face interface to carry out the score that recognition of face obtains to target facial image；
The corresponding similarity attenuation coefficient of the target user is calculated according to the ID card information；
Based on the recognition of face score and the similarity attenuation coefficient using the user gradation assessment models trained, obtain
The current user's level of the target user；
The corresponding goal task of the target user, the mesh are screened from current task set according to the current user's level
Mark task is the task that task grade meets preset condition；
The goal task is pushed into the terminal.
2. the method according to claim 1, wherein obtaining the mesh from the corresponding terminal of target user described
Before the recognition of face score and ID card information of marking user, comprising:
Obtain the facial image to be detected of user to be detected；
Characteristics of image is extracted from the facial image to be detected；
Described image feature is inputted to the In vivo detection model trained, it is pre- to obtain the corresponding living body probability of the user to be detected
Measured value；
It is tied when determining that the corresponding In vivo detection result of the user to be detected is preset for first according to the living body probabilistic forecasting value
When fruit, Xiang Suoshu terminal sends confirmation instruction, and the confirmation instruction is used to indicate the terminal for the facial image to be detected
It is determined as target facial image；
It is tied when determining that the corresponding In vivo detection result of the user to be detected is preset for second according to the living body probabilistic forecasting value
When fruit, Xiang Suoshu terminal sends warning message.
3. the method according to claim 1, wherein it is described according to the current user's level from current task
Before screening the corresponding goal task of the target user in set, comprising:
Obtain the corresponding task identification of each task in the current task set；
The corresponding task type of each task and contract information are searched according to the task identification；
Each task pair is obtained using the task grade evaluation model trained based on the task type and the contract information
The task grade answered.
4. according to the method described in claim 3, it is characterized in that, the generation step of the task grade evaluation model includes:
The first training sample set is obtained, it includes that historic task is corresponding that first training sample, which concentrates each first training sample,
Task type, contract information and the first markup information；
Determine the model structure information of initiating task grade evaluation model, and the initialization initiating task grade evaluation model
Model parameter；
Based in first training sample task type and contract information obtained using the initiating task grade evaluation model
To the corresponding task grade of first training sample；
Based on the difference between obtained task grade and first markup information, the initiating task grade assessment is adjusted
The model parameter of model obtains goal task grade evaluation model；
The goal task grade evaluation model is determined as to the task grade evaluation model trained.
5. the method according to claim 1, wherein the method also includes:
Task is completed in the history for obtaining the target user；
The corresponding task identification of task is completed according to the history and searches corresponding task scoring；
The current user's level of the target user is adjusted according to task scoring.
6. according to claim 1 to method described in 5 any one, which is characterized in that the life of the user gradation assessment models
Include: at step
The second training sample set is obtained, the second training sample of each of described second training sample concentration includes history target user
Corresponding recognition of face score, ID card information and the second markup information；
Determine the model structure information of initial user grade evaluation model, and the initialization initial user grade evaluation model
Model parameter；
The corresponding similarity decaying of the history target user is calculated according to the corresponding ID card information of the history target user
Coefficient；
Based on the corresponding recognition of face score of the history target user and similarity attenuation coefficient using described initial user etc.
Grade assessment models, obtain the corresponding user gradation of the history target user；
Based on the difference between the second markup information in obtained user gradation and the second training sample, adjust described initial
The model parameter of user gradation assessment models obtains target user's grade evaluation model；
Target user's grade evaluation model is determined as to the user gradation assessment models trained.
7. a kind of task driving means, which is characterized in that described device includes:
Data acquisition module, for obtaining the recognition of face score and identity of the target user from the corresponding terminal of target user
Information is demonstrate,proved, the recognition of face score is that the terminal calls Ministry of Public Security's recognition of face interface to carry out face to target facial image
Identify obtained score；
Similarity attenuation coefficient computing module, for calculating the corresponding similarity of the target user according to the ID card information
Attenuation coefficient；
Current user's level obtains module, has instructed for being used based on the recognition of face score and the similarity attenuation coefficient
Experienced user gradation assessment models, obtain the current user's level of the target user；
Goal task screening module, for screening the target user from current task set according to the current user's level
Corresponding goal task, the goal task are the task that task grade meets preset condition；
Goal task pushing module, for the goal task to be pushed to the terminal.
8. device according to claim 7, which is characterized in that described device further include: In vivo detection module, for obtaining
The facial image to be detected of user to be detected；Characteristics of image is extracted from the facial image to be detected；By described image feature
The In vivo detection model trained is inputted, the corresponding living body probabilistic forecasting value of the user to be detected is obtained；When according to the work
When body probabilistic forecasting value determines that the corresponding In vivo detection result of the user to be detected is the first default result, Xiang Suoshu terminal hair
Confirmation is sent to instruct, the confirmation instruction is used to indicate the terminal and the facial image to be detected is determined as target face figure
Picture；When determining that the corresponding In vivo detection result of the user to be detected is the second default result according to the living body probabilistic forecasting value
When, Xiang Suoshu terminal sends warning message.
9. a kind of computer equipment, including memory and processor, the memory are stored with computer program, feature exists
In the step of processor realizes any one of claims 1 to 6 the method when executing the computer program.
10. a kind of computer readable storage medium, is stored thereon with computer program, which is characterized in that the computer program
The step of method described in any one of claims 1 to 6 is realized when being executed by processor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811544266.7A CN109816200B (en) | 2018-12-17 | 2018-12-17 | Task pushing method, device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811544266.7A CN109816200B (en) | 2018-12-17 | 2018-12-17 | Task pushing method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109816200A true CN109816200A (en) | 2019-05-28 |
| CN109816200B CN109816200B (en) | 2023-11-28 |
Family
ID=66601988
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811544266.7A Active CN109816200B (en) | 2018-12-17 | 2018-12-17 | Task pushing method, device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109816200B (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298331A (en) * | 2019-07-05 | 2019-10-01 | 中国计量大学 | A kind of testimony of a witness comparison method |
| CN111259245A (en) * | 2020-01-16 | 2020-06-09 | 腾讯音乐娱乐科技（深圳）有限公司 | Work push method, device and storage medium |
| CN111275448A (en) * | 2020-02-22 | 2020-06-12 | 腾讯科技（深圳）有限公司 | Face data processing method and device and computer equipment |
| CN112084408A (en) * | 2020-09-08 | 2020-12-15 | 中国平安财产保险股份有限公司 | List data screening method and device, computer equipment and storage medium |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112949599A (en) * | 2021-04-07 | 2021-06-11 | 广州弘晟计算机系统有限公司 | Candidate content pushing method based on big data |
| CN113177480A (en) * | 2021-04-29 | 2021-07-27 | 上海商汤智能科技有限公司 | Financial business processing method, device, equipment and medium based on face recognition |
| CN113240335A (en) * | 2021-06-09 | 2021-08-10 | 武汉空心科技有限公司 | Engineer skill proficiency assessment method |
| CN116049560A (en) * | 2023-01-19 | 2023-05-02 | 阿里巴巴（中国）有限公司 | Information processing method and device |
| CN117437824A (en) * | 2023-12-13 | 2024-01-23 | 江西拓世智能科技股份有限公司 | Speech training methods and related devices |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654056A (en) * | 2015-12-31 | 2016-06-08 | 中国科学院深圳先进技术研究院 | Human face identifying method and device |
| CN107818313A (en) * | 2017-11-20 | 2018-03-20 | 腾讯科技（深圳）有限公司 | Vivo identification method, device, storage medium and computer equipment |
| CN107958317A (en) * | 2016-10-17 | 2018-04-24 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus that crowdsourcing participant is chosen in crowdsourcing project |
| CN108062538A (en) * | 2017-12-29 | 2018-05-22 | 成都智宝大数据科技有限公司 | Face identification method and device |
| CN108197557A (en) * | 2017-12-28 | 2018-06-22 | 深圳云天励飞技术有限公司 | Testimony of a witness consistency check method, terminal device and computer readable storage medium |
| CN108573054A (en) * | 2018-04-24 | 2018-09-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
| CN108764932A (en) * | 2018-04-09 | 2018-11-06 | 国网山东省电力公司 | Electricity business hall intelligence graded dispatching method and system based on recognition of face |
-
2018
- 2018-12-17 CN CN201811544266.7A patent/CN109816200B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105654056A (en) * | 2015-12-31 | 2016-06-08 | 中国科学院深圳先进技术研究院 | Human face identifying method and device |
| CN107958317A (en) * | 2016-10-17 | 2018-04-24 | 腾讯科技（深圳）有限公司 | A kind of method and apparatus that crowdsourcing participant is chosen in crowdsourcing project |
| CN107818313A (en) * | 2017-11-20 | 2018-03-20 | 腾讯科技（深圳）有限公司 | Vivo identification method, device, storage medium and computer equipment |
| CN108197557A (en) * | 2017-12-28 | 2018-06-22 | 深圳云天励飞技术有限公司 | Testimony of a witness consistency check method, terminal device and computer readable storage medium |
| CN108062538A (en) * | 2017-12-29 | 2018-05-22 | 成都智宝大数据科技有限公司 | Face identification method and device |
| CN108764932A (en) * | 2018-04-09 | 2018-11-06 | 国网山东省电力公司 | Electricity business hall intelligence graded dispatching method and system based on recognition of face |
| CN108573054A (en) * | 2018-04-24 | 2018-09-25 | 百度在线网络技术（北京）有限公司 | Method and apparatus for pushed information |
Cited By (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298331B (en) * | 2019-07-05 | 2021-04-06 | 中国计量大学 | Witness comparison method |
| CN110298331A (en) * | 2019-07-05 | 2019-10-01 | 中国计量大学 | A kind of testimony of a witness comparison method |
| CN111259245B (en) * | 2020-01-16 | 2023-05-02 | 腾讯音乐娱乐科技（深圳）有限公司 | Work pushing method, device and storage medium |
| CN111259245A (en) * | 2020-01-16 | 2020-06-09 | 腾讯音乐娱乐科技（深圳）有限公司 | Work push method, device and storage medium |
| CN111275448A (en) * | 2020-02-22 | 2020-06-12 | 腾讯科技（深圳）有限公司 | Face data processing method and device and computer equipment |
| CN112084408A (en) * | 2020-09-08 | 2020-12-15 | 中国平安财产保险股份有限公司 | List data screening method and device, computer equipment and storage medium |
| CN112084408B (en) * | 2020-09-08 | 2023-11-21 | 中国平安财产保险股份有限公司 | List data screening method, device, computer equipment and storage medium |
| CN112329638A (en) * | 2020-11-06 | 2021-02-05 | 上海优扬新媒信息技术有限公司 | Image scoring method, device and system |
| CN112949599A (en) * | 2021-04-07 | 2021-06-11 | 广州弘晟计算机系统有限公司 | Candidate content pushing method based on big data |
| CN113177480A (en) * | 2021-04-29 | 2021-07-27 | 上海商汤智能科技有限公司 | Financial business processing method, device, equipment and medium based on face recognition |
| CN113177480B (en) * | 2021-04-29 | 2024-09-17 | 上海商汤智能科技有限公司 | Financial business processing method, device, equipment and medium based on face recognition |
| CN113240335A (en) * | 2021-06-09 | 2021-08-10 | 武汉空心科技有限公司 | Engineer skill proficiency assessment method |
| CN116049560A (en) * | 2023-01-19 | 2023-05-02 | 阿里巴巴（中国）有限公司 | Information processing method and device |
| CN117437824A (en) * | 2023-12-13 | 2024-01-23 | 江西拓世智能科技股份有限公司 | Speech training methods and related devices |
| CN117437824B (en) * | 2023-12-13 | 2024-05-14 | 江西拓世智能科技股份有限公司 | Speech training method and related device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109816200B (en) | 2023-11-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109816200A (en) | Task method for pushing, device, computer equipment and storage medium | |
| CN110992167B (en) | Bank customer business intention recognition method and device | |
| CN109784654A (en) | Task creating method, device, computer equipment and storage medium | |
| CN109670437B (en) | Age estimation model training method, facial image recognition method and device | |
| CN112712429A (en) | Remittance service auditing method, remittance service auditing device, computer equipment and storage medium | |
| CN109858381A (en) | Biopsy method, device, computer equipment and storage medium | |
| CN111428448B (en) | Text generation method, device, computer equipment and readable storage medium | |
| CN109729383A (en) | Double record video quality detection methods, device, computer equipment and storage medium | |
| CN111736940B (en) | Business interface display method and device for intelligent counter | |
| CN111160275B (en) | Pedestrian re-recognition model training method, device, computer equipment and storage medium | |
| CN111597818B (en) | Call quality inspection method, device, computer equipment and computer readable storage medium | |
| CN109255747A (en) | A kind of intelligent checks method that information is declared | |
| CN117036843B (en) | Target detection model training method, target detection method and device | |
| CN116958846A (en) | Video detection methods, devices, equipment, media and products | |
| CN116205726B (en) | Loan risk prediction method and device, electronic equipment and storage medium | |
| CN114067401A (en) | Target detection model training and identity verification method and device | |
| CN113902041A (en) | Method and device for training and authentication of target detection model | |
| CN114692785B (en) | Behavior classification method, device, equipment and storage medium | |
| CN109766772A (en) | Risk control method, device, computer equipment and storage medium | |
| US20160239559A1 (en) | Document classification system, document classification method, and document classification program | |
| CN112329735B (en) | Training method of face recognition model and online education system | |
| CN113239233A (en) | Personalized image recommendation method, device, equipment and storage medium | |
| CN109711301A (en) | Driver review method, apparatus, computer equipment and storage medium | |
| CN116189063A (en) | Key frame optimization method and device for intelligent video monitoring | |
| CN117808573A (en) | A risk level identification method and device based on imbalanced data sets |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |